Synthesis of novel conjugates based on a functionalized cyclo[DKP-isoDGR] integrin ligand and potent cytotoxic agents

(1)

UNIVERSITA DEGLI STUDI DELL'INSUBRIA

Dipartimento di Scienza e Alta Tecnologia PhD course in Chemical Sciences, cycle XXX

EuropeanTraining Network MAGICBULLET

Synthesis of novel conjugates based on a functionalized cyclo[DKP-isoDGR] integrin ligand and potent cytotoxic

agents

Lizeth Alicia Bodero Padilla

Tutor: Prof. Umberto Piarulli – Università degli Studi dell’Insubria

Academic Co-Tutor: Prof. Cesare Gennari – Università degli Studi di Milano Industrial Co-Tutor: Dr. Christoph Müller - Heidelberg Pharma

Academic Year: 2017/2018

(2)

1 Doctoral Final Examination: November !5^th, 2018

Examination Committee: Prof. Ines Neundorf Prof. Norbert Sewald Prof. Gianluigi Broggini

(3)

2 The work herein described was performed at the Insubria University at the Department of Science and Hight Technology – Laboratory of Organic Chemistry in the period from July 2015 to June 2018 under the supervision of Prof. Umberto Piarulli. This project has received funding from the European Union’s Horizon 2020 research and innovation program under the Marie Sklodowska-Curie grant agreement No 642004.

I gratefully acknowledge my supervisor Prof. Umberto Piarulli for the opportunity of being part of this multidisciplinary project, for his permanent support and for trusting in my work. I deeply thank Prof. Cesare Gennari and Dr. Müller for their support during my secondments and for their important contribution to the development of my project. I thank to my colleagues in Como: Sara, Clem, Silvia, Bob, Pippo, Mirko and Luca for their friendship, constant help and for creating a great work environment. Finally, I would like to thank to all the MAGICBULLET network: Marcel, Norbert, Gàbor, Ines, Pirjo, Cesare, József, Christoph, Christian, Hans, Ralph, Paula, Adina, Eduard, Abi, Lucy, Francy, Barbara, J.K., Ana, Andre, Paula, Clem, Sabine, Andrea and Ivan for these three years of intensive collaboration and beautiful experiences.

This work is dedicated to my dear family: David, Alicia, Maybe, Davicho and Yeshi.

(4)

3

Abbreviations

5-FU: 5-Fluoracil

6-MP: 6-Mercaptopurine

A549: Human lung carcinoma cell line ACN: Acetonitrile

ACPP: Active cell penetrating peptide ADC: Antibody-drug conjugate ADCC: Antibody-dependent cellular cytotoxicity

AKt/PKB: Protein kinase B Ala: Alanine

ALL: Acute lymphoblastic leukemia

AMAS: N-(α-maleimidoacetoxy) succinimide ester

AML: Acute myeloid leukemia APN: Aminopeptidase N aq.: Aqueous solution Bn: Benzyl

Boc: tert-Butyloxycarbonyl CAIX: Carbonic anhydrase IX

CCRF-CEM: human leukemic lymphoblasts CDC: Complement dependent cytotoxicity Cit: Citrulline

CPT: Camptothecin

CuAAC: Copper (I) catalyzed alkyne-azide cycloaddition

DAR: Drug-antibody ratio

DAVBH: Desacetyl vinblastine hydrazide DCC: N,N'-Dicyclohexylcarbodiimide DCM: Dichloromethane

DIC: N,N'-Diisopropylcarbodiimide DIPEA: N,N-Diisopropylethylamine DKP Diketopiperazine

DM1: N2'-deacetyl-N2'-(3-mercapto-1- oxopropyl)-maytansine

DMAP: 4-Dimethylaminopyridine

DMF: N,N-Dimethylformamide DMSO: Dimethyl sulfoxide DNA: Deoxyribonucleic acid DOX: Doxorubicin

DOXSF: Doxsaliform

DUPA: 2-[3-(1,3-dicarboxypropyl)ureido]

pentanedioic acid ECM: Extracellular matrix

EDC: N-(3-Dimethylaminopropyl)-Nʹ- ethylcarbodiimide hydrochloride EDT: ethanedithiol

EEDQ: N-Ethoxycarbonyl-2-ethoxy-1,2- dihydroquinoline

EGFR: Epithelial growth factor receptor Ep-CAM: Epithelial cell adhesion molecule eq.: equivalents

ESI Electrospray ionization EtOAc: Ethyl acetate

FACS: Fluorescence-activated cell sorting FAK: Focal adhesion kinases

FDA: US Food and Drug Administration FL: Follicular lymphoma

Fmoc: 9-Fluorenylmethoxycarbonyl FN: Fibronectin

FR: Folate receptor GI: Gastrointestinal

GM: Glioblastoma multiforme

HAMA: Human anti-mouse antibodies HATU: O-(7-azabenzotriazol-1-yl)-

tetramethyl-uronium hexafluorophosphate HOAt 1-Hydroxy-7-azabenzotriazole

HPLC: High performance liquid chromatography

HT29: Human colon cancer

HUVEC: Human umbilical endothelial cells

(7)

6 IC50: Maximum half inhibitory concentration

ICAM 1: (Intercellular Adhesion Molecule 1 IgG: Immunoglobulin G

IL-8: Interleukin 8 iPr: Isopropyl

isoDGR: isoAspartic-Glycine-Arginine LNCaP: androgen-sensitive human prostate adenocarcinoma cell line

LRP1: Low-density lipoprotein LTT: Ligand-targeted therapeutics mAb: Monoclonal antibody MS: Mass spectroscopy

mCRC: metastatic colorectal cancer

MCRPC: Metastatic castration-resistant prostate cancer

MDA-MB-468: Human breast cancer cell line

MED: Minimum effective dose MeOH: Methanol

MIDAS: Metal ion-dependent adhesion site MMAE: Monomethyl auristatin E

MMAF: Monomethyl auristatin F MMP: Matrix metalloproteinase MTD: Maximum tolerated dose

Mtr: 4-methoxy-2,3,6-trimethyl benzene sulphonyl

MTT: 3-(4,5-dimethylthiazol-2-yl)-2,5- diphenyltetrazolium bromide

MW Molecular weight NaHCO3: Sodium bicarbonate NGR: Asparagine-Glycine-Arginine NHL: Non-Hodgkin’s lymphoma NHS: N-Hydroxysuccinimide NMR: Nuclear Magnetic Resonance NSCLC: Non-small-cell lung cancer OATP: organic anion transporting polypeptide

PABA: p-aminobenzyl alcohol

PABC: p-aminobenzyl carbamate PaCa2: human pancreatic duct adenocarcinoma

PBS Phosphate-buffered saline PEG: Polyethylene glycol

pHLIP: pH low insertion peptides PNP: p-nitrophenyl chloroformate

PSMA: Prostate specific membrane antigen PTX: Paclitaxel

RGD: Arginine-Glycine-Aspartic rt.: room temperature

SCLC: Small-cell lung cancer SIP: Small immune proteins

SMDC: Small molecule-drug conjugate SPDP: N-Succinimidyl 3-(2-pyridyldithio) propionate

SSTR: Somatostatin receptor

sulfo-Lc-SMPT : sulfosuccinimidyl 6-[α- methyl-α-(2-pyridyldithio)toluamido]

hexanoate

TFA: Trifluoroacetic acid THF: Tetrahydrofuran TK: Tyrosine kinase TMS Tetramethylsilane TNF: Tumor necrosis factor tR: Retention time

U87-MG: Human glioblastoma cell line Val: Valine

VEGF: Vascular endothelial growth factor receptor

(8)

7

Project overview

The lack of selectivity is one of the main limitations of traditional chemotherapy because of the severe side effects associated to high drug dosages. Targeted drug delivery is therefore a growing-interest field in cancer therapy as a strategy for overcoming the systemic cytotoxicity. This approach is inspired by the “magic-bullet”

concept of Paul Ehrlich, awarded with the Medicine Nobel Prize in 1908 for his work in immunotherapy, concept applied to cancer therapy nowadays to propose the use of drug-delivery vehicles (monoclonal antibodies, peptides, nanoparticles, polymers) targeting a specific antigen or receptor to liberate the payload at the tumor site without affecting the healthy tissue.

The antibody-drug conjugates (ADCs) and the small molecule-drug conjugates (SMDC) belong to this new generation of therapeutics. In ADCs, the targeting agent is a monoclonal antibody (mAb) while in SMDCs the targeting is performed by a low molecular weight ligand (peptide, vitamin or peptidomimetic). In both cases the targeting moiety is connected to a potent warhead by means of a stable linker and they are expected to efficiently deliver the cytotoxic agent to the tumor cells preferentially via receptor-mediated endocytosis. Currently, four ADCs: Mylotarg, Kadcyla, Adcetris and Besponsa, have been approved by the US FDA for the treatment of different cancers. Despite this success, ACDs present some drawbacks related to the use of mAbs such as high manufacturing costs, unfavorable pharmacokinetics (low tissue diffusion and low accumulation rate) and possible immune response, for this reason smaller formats like SMDCs have become an interesting alternative for the selective delivery of drugs into tumors.

Our research group has developed during the last decade a number of cyclic peptidomimetic ligands containing the tripeptide RGD or isoDGR sequences and the bifunctional diketopiperazine (DKP) scaffold. These ligands are recognized by the integrin receptor αVβ3, which is widely expressed on the blood vessels of several human

(9)

8 cancers (e.g. breast cancer, glioblastoma, pancreatic tumor, prostate carcinoma) but not on the vasculature of healthy tissues, constituting a suitable therapeutic target. In particular, the cyclo[DKP-isoDGR] integrin ligand has shown not only a high binding affinity and selectivity for the purified receptor αVβ3 but also an integrin antagonist activity, becoming a promising ligand for the preparation of SMDCs.

This PhD thesis describes the synthesis and biological evaluation of SMDCs containing the functionalized cyclo[DKP-isoDGR] integrin ligand and potent cytotoxic drugs (α- amanitin, MMAE and MMAF) combined via different linkers and spacers. The purpose of this project is to study the efficacy of the cyclo[DKP-isoDGR] integrin ligand developed by our research group as a vector for targeted drug delivery. The work is divided in three chapters: the Chapter 1 introduces the definition of SMDC as part of the new targeted therapies strategies; the Chapter 2 presents the characteristics of the integrin receptors family, previous work on targeted drug-delivery via integrin ligands and the synthesis of the functionalized cyclo[DKP-isoDGR] integrin ligand; and the Chapter 3 presents the synthesis of the SMDCs containing the cyclo[DKP-isoDGR]

integrin ligand and different cytotoxic agents, along with the discussion of the in vitro results, including binding affinity tests towards the isolated αVβ3 integrin receptor and antiproliferative activity assays in cancer cell lines with different levels of αVβ3

expression.

Finally, the experimental procedures and in vitro test protocols are detailed in the Experimental Section, together with the spectroscopic and analytical data of the new products.

(10)

9

Chapter 1. Targeted drug delivery for tumor therapy

1.1. Targeted cancer therapies

Cancer has a major impact on society as it represents one of the main causes of death worldwide. Statistics from the World Health Organization indicate that 8.8 million people died of cancer in 2015 and, approximately 1 in 6 deaths globally is related to this desease.¹

Among the anticancer therapies developed during the last decades, chemotherapy remains the most employed together with radiation therapy and surgery. Traditional chemotherapy uses low-molecular weight drugs (Figure 1) that modify or interrupt the cell cycle at different stages.^2,3 These cytotoxic drugs can be classified by their mechanism of action in:

- Antimetabolites

These drugs can alter essential biological pathways by mimic nitrogenous bases or inhibiting enzymes involved in the synthesis of nucleic acids. 5-Fluoracil (5-FU) and 6-mercaptopurine (6-MP), analogues of pyrimidine and purine respectively, are examples of antimetabolite drugs. The incorporation of these analogues during the phase S of cell cycle interrupts the replication of DNA and leads to apoptosis.

Another example of this group is methotrexate, an antifolate that blocks the synthesis of nucleotides by inhibition of the dihydrofolate reductase.⁴

- DNA damaging agents

- Alkylating agents: They act by alkylating DNA on purine bases blocking replication.

Nitrogen mustards derivatives (e.g. cyclophosphamide, chlorambucil, melphalan);

nitrosoureas (e.g. carmustine, lomustine, semustine); triazenes (e.g. dacarbazine,

(11)

10 temozolomide) and natural products like mitomycin C and streptozotocin belong to this group.^4,5

- Cross-linking agents: Some alkylating agents can also bind DNA causing inter-strand crosslink (DNA crosslink on opposite strands of the double helix), with subsequent double toxic effect in cells. Platinum complexes and derivates (e.g. cisplatin, carboplatin, oxaliplatin) can form intra-strand crosslinks when forming adducts with adjacent bases on the same DNA strand. Both inter-strand and intra-strand crosslinks lead to apoptosis by interruption of DNA replication.⁵

- Intercalating agents: Bind between base pairs of nucleic acids preventing replication. Examples of this group of drugs are the anthracyclines doxorubicin, daunorubicin, epirubicin, mitoxantrone and antinomycin-D.^4,5

- Toposisomerase poisons: Topoisomerases are enzymes responsible for the cleavage, annealing and topological state of DNA double helix. Toposisomerase I inhibitors (e.g. camptothecin, irinotecan, topotecan) and toposisomerase II inhibitors (etoposide, anthracyclines) trap the DNA-enzyme complex inhibiting replication fork progression.⁶

- Antitubulin agents

Tubulin is a globular protein that plays an essential role in cellular replication.

Antitubulin agents, also known as mitotic inhibitors, alter the microtubule polymerization dynamics (coexistence of tubulin assembly and disassembly) blocking the division of the nucleus and leading cell to apoptosis. There exist two groups of antitubulin agents: microtubules stabilizers (e.g. Paclitaxel, docetaxel, epothilones) and microtubule inhibitors (e.g. vincristine, vinblastine, dolastatin 10, colchicine).⁷

(12)

11 Figure 1. Molecular structures and use of common anticancer drugs⁸

The cytotoxic agents described above are expected to attack preferentially the tumor cells, as these undergo much more rapid proliferation than normal cells. Unfortunately, these drugs can also kill normal dividing cells in the body (e.g. hair, bone marrow, gastrointestinal track) and accumulate in other organs rather than tumor area (Figure 2), displaying severe side effects. Due to the systemic cytotoxicity, the administrated dose is often reduced to suboptimal level with poor benefit for the patient.^3,9,10

Figure 2. Biodistribution of ¹¹C-docetaxel in male patient with metastatic malignant pleural mesothelioma. PET scans at different time points (0–6, 8–19, 23–39 and 42–63 min) display low drug uptake in the affected region (pleural mesothelium) but high accumulation in the liver and intestine.

Adapted with permission from A.A.M. van der Veldt, N.H. Hendrikse, E.F. Smit et al. Eur. J. Nucl. Med.

(13)

12 With the purpose of improving the therapeutic index (maximum tolerated dose/minimum effective dose) of anticancer drugs, modern strategies are oriented towards targeted therapies. These are based on the concept of “magic-bullet”, envisaged by Paul Ehrlich (Medicine Nobel Prize, 1908) who coined the term referring to a therapeutic agent that could attack the bacteria responsible of diseases without hurting the host.¹²

Figure 3. Targeted chemotherapy approach. Adapted with permission from Siler Panowski, Sunil Bhakta, Helga Raab, Paul Polakis and Jagath R. Junutula; mAbs, 2014, 6 (1), 34-45. Copyright © 2014 Landes Bioscience

Nowadays, targeted cancer therapies aim at selectively killing cancer cells by interfering with essential pathways involved in tumor growth (e.g. signal-transduction pathways) or by efficiently delivering the cytotoxic agents to the tumor without compromising the healthy cells. Most common targets are molecular markers or antigens that play an important role in cell proliferation (e.g. cell surface proteins, glycoproteins, or carbohydrates) and that are over-expressed in tumor cells compared with normal tissues.^3,13

Three main approaches can be identified in this field: small molecules tyrosine kinase inhibitors designed to prevent the activation of signaling pathways dysregulated in tumor cells (e.g. imatinib – Gleevec®, sunitinib)^14–16; monoclonal antibodies (mAbs) targeting specific antigens displayed in tumor cells; and ligand-targeted therapeutics (LTT) where a drug-delivery vehicle (e.g. monoclonal antibody, small ligand, peptide, nanoparticle, polymer) is used to target specific antigens or receptors and liberate the payload at the tumor site.^17,18 Antibody-drug conjugates (ADCs) and small molecule-

(14)

13 drug conjugates (SMDCs) belong to this last category and will be described in the next sections.

1.2. Monoclonal antibodies

Antibodies, also known as immunoglobulins (Ig), are a group of glycoproteins produced by the immune system that detect and selectively bind antigens displayed in abnormal cells, prompting an immune attack that destroys the cell expressing the antigen.¹⁹ In 1975, Köhler and Milstein developed the hybridoma technology,²⁰ where antibodies produced by B lymphocytes of mice were isolated and fused with immortal myeloma cell lines to obtain clonal cells known as hybridomas. Hybridomas can be cultured to produce large amounts of identical antibodies specific for an antigen, these are called monoclonal antibodies (mAbs).

Figure 4. Structure and classification of monoclonal antibodies (mAbs).²¹ Fab: fragment antigen-binding;

Fc: fragment crystallizable; CDR: complementary determining region; A) murine antibody; B) Chimeric antibody: murine variable regions and human constant regions; C) Humanized antibody: human variable and constant region, murine CDRs; D) Fully human antibody. Adapted with permission from K. R.

The therapeutic potential of mouse mAbs (Figure 4A) was restricted by the response of patients’ immune system, which recognized the antibodies as foreign entities and generated human anti-mouse antibodies (HAMA), resulting in rapid clearance of the mAb from circulation. Further advances in recombinant DNA technology led to the production of chimeric mAbs (Figure 4B) where the constant regions of the murine antibody were replaced by human constant regions sequences, retaining the murine variable domains responsible for antigen binding. Later, it was also possible to replace the variable sequences of mouse mAbs with human sequences to obtain humanized mAbs with less than 10% mouse protein. More recently, the phage display technology

(15)

14 and the use of transgenic mice have made possible the generation of fully human mAbs, significantly reducing the immune response reported for the first mAbs.^3,21,22

The antitumoral activity of monoclonal antibodies can be attributed to different cell- killing mechanisms.²³ Among them we can summarize:

- Direct action of the antibody: by binding the targeted receptor and displaying an antagonist activity, blocking the dimerization, kinase activation and downstream signaling, inhibiting proliferation and inducing apoptosis.

- Immune-mediated cell killing: induction of phagocytosis, complement dependent cytotoxicity (CDC), antibody-dependent cellular cytotoxicity (ADCC) or regulation of T-cell function.

- Effect of the antibody on tumor vasculature and stroma: vasculature receptor antagonism, stromal cell inhibition.

These mechanisms have been validated in the clinic, leading to the approval of more than 24 mAbs targeting 16 different antigens, that are currently available in the market for the treatment of an increasing number of cancers.²⁴ Some examples include the chimeric mAb Rituximab (the first mAb approved by the FDA for cancer treatment in 1997), that binds the CD20 antigen expressed on the surface of B cells, indicated for the treatment of follicular lymphoma (FL) and low-grade non-Hodgkin’s lymphoma (NHL);²⁵ Alemtuzumab, a humanized mAb that targets CD52 antigen overexpressed on malignant lymphocytes, approved for therapy of resistant lymphocytic leukemia;²⁶ Bevacizumab, another humanized mAb that binds to the vascular endothelial growth factor receptor (VEGF) on cancer cells, inhibiting the formation and growth of tumor blood vessels, used in the treatment of metastatic colon and kidney cancer, non-small cell lung cancer and glioblastoma;²⁷ the human mAb panitumumab, indicated for the treatment of metastatic colorectal cancer (mCRC) expressing the epithelial growth factor receptor (EGFR);²⁸ and the humanized mAb trastuzumab (Herceptin®), that targets the human epidermal growth factor receptor 2 (HER2) overexpressed in 20-30%

of breast cancers and some metastatic gastrointestinal (GI) cancers.^29,30

(16)

15

1.3. Antibody-drug conjugates (ADC)

The high specificity of monoclonal antibodies and the positive results obtained from the combination of mAbs with chemotherapy led to the idea of developing new conjugated entities where the mAbs were covalently bound to cytotoxic drugs. The so-called antibody-drug conjugates (ADCs) consist of a monoclonal antibody connected to a potent cytotoxic drug via a chemically stable linker to prevent the premature release of the payload in the blood circulation. Ideally, an ADC should liberate the cytotoxic agent at the tumor site after selectively binding its target, expressed on tumor cells surface, leaving the healthy cells unharmed.^13,31

The first generation of ADCs, based on chimeric or humanized mAbs and regular- potency cytotoxic payloads (e.g. doxorubicin, methotrexate), faced several problems in clinical trials because of immunogenicity, limited potency and insufficient selectivity.³² Learning from the early results, scientists optimized the ADCs design by selecting more specific targets, replacing chimeric mAbs by humanized or fully human mAbs to prevent immunogenicity, and using ultra-potent cytotoxic drugs (100-1000 times more potent).

The drug-antibody ratio (DAR) in the second-generation ADCs is around 4:1, resulting in more efficient cytotoxicity.^33,34 Currently, a new generation of ADCs is being developed with the incorporation of bispecific antibodies and site-specific conjugation of the drug that allows a better control of the DAR.³⁵

At present, four ADCs have received FDA approval and more than 60 are being evaluated at different stages of clinical trials (Figure 5). The first ADC commercialized was gemtuzumab ozogamicin (Mylotarg®, Wyeth-Pfizer), a first-generation ADC approved in 2000 for the treatment of CD33-positive acute myeloid leukemia (AML). It was voluntarily withdrawn from the US market in June 2010 and reintroduced in September 2017. The other three are second-generation ADCs: brentuximab vedotin (Adcetris®, Seattle Genetics) approved in 2011 for the treatment of anaplastic large cell lymphoma and Hodgkin lymphoma; ado-trastuzumab emtansine (Kadcyla®, Genentech), approved in 2013 for the treatment of HER2-positive breast cancer and inotuzumab ozogamicin (Besponsa®, Pfizer) approved in 2017 for the treatment of adults with relapsed or refractory B-cell precursor acute lymphoblastic leukemia (ALL).^31,36

(17)

16 Figure 5. Structure and classification of three ADCs approved by FDA.³⁵ A) The first-generation ADC gemtuzumab ozogamicin (Mylotarg®) contains a humanized mAb (IgG4) specific for CD33 antigen conjugated to 2–3 calicheamicin moieties which are attached via cleavable hydrazone linkers to random lysine residues; B) The second-generation ADC trastuzumab emtansine (Kadcyla ®) contains a humanized mAb (IgG1) specific for human epidermal growth factor receptor 2 (HER2) and 3–4 DM1 moieties attached via non-cleavable thioether linkers to random lysine residues; C) The second-generation ADC brentuximab vedotin (Adcetris®) contains a chimeric mAb (IgG1) specific for CD30 antigen and 4 monomethyl auristatin E (MMAE) moieties attached to the hinge region through a protease-cleavable Val-Cit linker. Adapted with permission from A. Beck; L. Goetsch; C. Dumontet; N. Corvaïa. Nat. Rev.

1.3.1. Design and mechanism of action of ADCs

Among the factors to consider for the development of efficient ADCs, the most determinant are the choice of the target, the binding affinity and immunogenicity of the antibody, the nature of the linker and the potency of the cytotoxic drug.³¹ Most ADCs are designed to kill cancer cells in a target-dependent mechanism that involves the internalization of the ADC via a receptor-mediated endocytosis pathway (Figure 6).^37,38 The first step in this process is the binding of the antibody to its antigen, localized preferentially on the cell surface of tumor cells. Once the ADC-antigen complex is formed, it is internalized into endosomes that subsequently mature and fuse with lysosomes. In the lysosomes, the drug is released via cleavage of the linker by specific proteases such as cathepsin B or by degradation of the ADC, then the free drug reaches its target in cytoplasm leading to cell death. The cell-killing mechanism depends of the class of cytotoxic drug used (e.g. tubulin polymerization inhibition by maytansines and auristatins, DNA damage by calcheamicins and duocarmycins).^34,38 Neighboring cancer

(18)

17 cells can also be killed when the free drug crosses the plasma membrane and access the extracellular environment in a process known as the bystander killing effect.^39,40

Figure 6. Design of ADCs and cell-killing mechanism. A) General structure of ADCs and factors to be consider for the design. B) Receptor-mediated endocytosis pathway: 1) the antibody moiety binds to its cell-surface antigen receptor target and form an ADC-antigen complex; 2) the complex is internalized into endosome that fuse with lysosomes; 3) the internalized complex undergoes lysosomal processing;

4) the cytotoxic payload is released inside the cytosol; 5) the payload reaches its target leading to cell death. Adapted with permission from Siler Panowski, Sunil Bhakta, Helga Raab, Paul Polakis and Jagath R. Junutula; mAbs, 2014, 6 (1), 34-45. Copyright © 2014 Landes Bioscience

Recent approaches in the design of ADCs have questioned the internalization requirement for ADC efficiency. In effect, potent activity in tumor preclinical models has been reported for non-internalizing ADCs directed against splice isoforms of fibronectin and tenascin-C, both expressed on the extracellular matrix of tumor blood vessels.^41–43 In this case, the release of the cytotoxic drug is triggered by the glutathione or proteases present in the extracellular space upon tumor cell death, followed by the passive diffusion of the drug, which should be lipophilic enough to guarantee a homogeneous drug delivery to the tumor.

1.3.2. Limitations of ADCs

Despite the successful approval of four ADCs and the remarkable progress achieved in this field, there remain some limitations concerning the immunogenicity of the antibody, the stability of the linker, the antigen targeting and the heterogeneity of the antigen expression in the tumor.^44,45 Specifically, in the case of solid tumors, the number of targeted receptors that ensure the internalization is relatively limited, and

(19)

18 contrary to hematologic tumors, the targets are overexpressed in a small portion of the patient populations (e.g. only 20% of breast cancer express HER2 and is eligible for the treatment with Kadcyla). Also, due to its large size antibodies do not extravasate and diffuse efficiently into tissue, once they reach the blood vessels they are trapped by the antigens located on perivascular tumor cells, preventing the targeting on the integrity of the tumor area. This is known as the “antigen-barrier” effect.^9,46

Figure 7. Various sizes antibody formats and alternative scaffolds for drug delivery.⁴⁷ A) Engineered antibody formats: (a) Immunoglobulin-G, (b) Small immune protein SIP, (c) diabody, (d) Fab fragment, (e) single chain Fv (scFv), (f) domain antibody (dAb); B) Protein scaffolds: (g) Designed ankyrin repeat protein (DARPin), (h) Adnectin,(monobody) (i) affibody, (j) knottin peptide, (k) bicyclic peptide. Adapted with permission from M. Deonarain; G. Yahioglu; I. Stamati et al. Antibodies 2018, 7 (2), 16.

Current strategies in ADC technology seek to overcome the pharmacokinetic limitations by using smaller formats such as antibody fragments, diabodies, mini-antibodies or small immune proteins (SIP) (Figure 7). Even though immunoglobulins (IgG) have a longer circulation half-life that allows maximal accumulation at the tumor site, recent studies support the concept that smaller formats have higher diffusion and extravasation coefficients, hence they can penetrate better the solid tumors.^48,49 Lower plasma exposure also reduces the risk of premature release of the payload and the lack of a Fc domain in mAb fragments can minimize cross-reactivity with Fc-receptors on various normal cells, reducing off-target cytotoxicity.^35,47

(20)

19

1.4. Small molecule-drug conjugates (SMDCs)

The search for smaller vehicles for targeted drug delivery generated a new class of ligand-targeted cytotoxic agents where the targeting moiety is a low-molecular weight ligand (e.g. peptide, vitamin, peptidomimetic) with potentially favorable pharmacokinetic properties and that can be easily accessed by chemical synthesis. The so called small-molecule drug conjugates (SMDC) are similar to ADCs in structure and mechanism of action. They contain a linker system that connects the ligand to the cytotoxic drug and are expected to deliver the payload at its intracellular target in the tumor by receptor-mediated endocytosis (Figure 8). The considerations for the design of efficient ADCs are also valid for the SMDCs, being most determinant the choice of the target, the ligand, the nature of the linker and the potency of the payload.^9,18,50

Figure 8. General structure and mechanism of action of SMDCs. A) General structure of SMDCs. B) Receptor-mediated endocytosis pathway: (a) the targeting moiety binds to its cell-surface receptor target and form a SMDC-receptor complex; (b) the complex is internalized into endosomes that fuse with lysosomes; (c) the internalized complex undergoes lysosomal processing; the linker is cleaved and the cytotoxic drug is released to reaches its intracellular target; (d) the receptor is recycled to cell surface. Adapted from https://endocyte.com/.

1.4.1. Target selection

One of the most important aspects for the choice of a suitable target is the receptor expression profile. This includes the expression of the targeted receptor in tumor cells vs normal tissues and the absolute level of receptor’s isoforms expression in tumor cells. Ideally, the targeted receptor must be sufficiently overexpressed in tumor cells compared to normal to avoid off-target cytotoxicity; some studies have stablished as acceptable a 3-fold or higher magnitude of receptor overexpression in cancer cells.^51,52 For example, the expression of the folate receptor type α (FRα) is about 2.8 million

(21)

20 receptors per cancer cell⁵³ and the prostate-specific membrane antigen (PSMA) is expressed in 1 million excess by LNCaP prostate cancer cell line⁵⁴. Other receptors overexpressed in a variety of cancer cells are the somatostatin receptor 2 (SSTR2), the sigma non-opioid intracellular receptor (SIGMAR1 and 2), cell-adhesion proteins ICAM1, LFA1 and CD24 and certain integrins.^9,13

Most targeted receptors are expressed on the surface of cancer cells, allowing a better accessibility for the targeting ligand. In general, once the ligand binds the cell-surface receptor, forms the SMDC-receptor complex that is internalized via endocytosis to an intracellular compartment (recycling endosome or a lysosome) where the complex is dissociated, allowing the receptor to be either degraded or recycled to the cell surface (Figure 8B). As the availability of empty receptors on the targeted tumor cell depends on the rate of return of unoccupied receptors, an ideal receptor will be frequently recycled or resynthesized following degradation.^9,50

1.4.2. Choice of the ligand

The binding affinity and specificity of the small ligand for the targeted receptor are basic for the optimal performance of SMDCs. A high binding affinity allowing the access to the targeted receptor and rapid extravasation, can increase the accumulation ratio tumor:blood, tumor:organ of the drug, avoiding the fast clearance often associated to the use of small molecules.^13,18 The specificity is also very important, especially when there are other members of the receptor family that can be recognized by the ligand, compromising the tumor targeting. Some of the ligands currently used in SMDCs include PSMA ligands,^50,55 folic acid analogues⁵⁶ and carbonic anhydrase IX (CAIX) ligands (Figure 9).^57,58

(22)

21 Figure 9. Small ligands targeting PSMA, folate receptor and CAIX

Another important aspect is the derivatizability of the ligand. In order to be conjugated to the linker (or to the connecting spacer) the targeting ligand should preferably have a derivatizable functional group (e.g. carboxylic acid, amine, alcohol, thiol) that enable the coupling to further entities via simple chemistry (e.g. formation of amides, carbamates, oximes, esters, carbonates or disulphides). One advantage of the use of small molecule ligands is that they can be prepared by chemical synthesis and optimized through structure-activity relationship (SAR) studies to identify the sites where modification will not interfere with receptor binding.^50,59

1.4.3. Linker design

The design of the linker system is a key factor in the optimization of the SMDCs because it has direct influence in the pharmacokinetic profile. As most SMDCs are designed to be cleaved or degraded intracellularly, the linker should be stable at physiological conditions but assure an efficient release of the payload after receptor-mediated internalization. In general, linkers can be divided in:

- Acid-sensitive linkers: functional groups (e.g. esters or hydrazone) that remain stable in blood circulation (pH 7.5) and get hydrolyzed in acidic tumor micro- environment (lysosomal pH 4.8 and endosomal pH 5–6). The hydrazone linker has been used in conjugates containing doxorubicin,^60–62 paclitaxel⁶³ and Pt agents.^64,65

(23)

22 This linker is applicable for conjugating drugs or their derivatives containing a chemical moiety (e.g. ketone or aldehyde) that can be coupled to a hydrazine- terminated linker (Figure 10).⁶⁶

Figure 10. Structures of hydrazone-based linker types as applied for doxorubicin conjugation: (a) acyl hydrazone, (b) alkoxycarbonyl hydrazone, (c) sulfonyl hydrazone. Doxorubicin release occurs primarily in acidic compartments such as endosomes and lysosomes as a result of acid-catalyzed hydrolysis of the hydrazone linker. Adapted with permission from P. T. Wong and S. K. Choi; Chem. Rev. 2015, 115, 3388- 3432. Copyright © 2015 American Chemical Society.

- Disulfide linkers: disulfide bonds can be reduced inside cytoplasm by endogenous thiol molecules (e.g. cysteine, glutathione). Glutathione is a low molecular weight thiol present in the cytoplasm (0.5–10 mM) and the extracellular environment in minor scale (2–20 µM in plasma). Tumor cells present elevated levels of glutathione (10-20 mM) due to stress conditions such as hypoxia.^67,68 This higher expression serves as a mechanism for the controlled release of the drug into the targeted cells (Figure 11).

(24)

23 Figure 11. Mechanism of drug release of self-immolative disulfide linkers. (a) when reduced by glutathione, a self-immolative cleavage of the disulfide bond leads to release of a free drug with the formation of stable byproducts (a) CO2 + a thiirane, (b) CO2 + thiolactone, (c) CO2 + thioquinone methide. GSH = glutathione; X = O or NH. Adapted with permission from: M. Srinivasarao and P. S. Low;

- Enzymatically-cleavable linkers: these linkers are generally constituted by short peptide sequences (e.g. Val-Ala, Val-Cit) designed to be cleaved by enzymes upregulated or activated inside the tumor cells, such as the lysosomal protease cathepsin B. In serum conditions (pH 7.5) these proteases are inactivated due to the presence of different protease inhibitors, for this reason the peptide linker is stable in systemic circulation and it is only cleaved upon internalization in tumors. In 2011, the FDA approved the ADC Adcetris® containing a Val-Cit linker connected to the self-immolative p-aminobenzylcarbamate-monomethyl auristatin E and an anti- CD30-mAb.^34,69

Figure 12. Enzymatic cleavage and drug release of paclitaxel prodrug containing the peptide linker [D]-Ala–Phe–Lys. Prodrug [D]-Ala-Phe-Lys-PABC-PTX a undergoes enzymatic cleavage giving b, 1,6- elimination of self-immolative p-aminobenzylcarbamate moiety gives the metabolite c. Finally, intramolecular cyclization allows the release of the paclitaxel.⁷⁰ Adapted with permission from A. dal

(25)

- Non-cleavable linkers: alkyl or polymeric moieties that liberate the payload only after lysosomal degradation of the conjugate inside the cell. Their main advantage is the increased plasma stability compared to cleavable linkers⁷¹ and the specificity of the drug release mechanism. This type of linker has been successfully used in ADCs, notably in trastuzumab emtansine (Kadcyla®), that contains a non-reducible thioether, N-succinimidyl-4-(N-maleimidomethyl) cyclohexane-1-carboxylate (SMCC or MC after conjugation), that connects the antibody with the maytansinoid DM1.⁷² 1.4.4. Payload selection

The choice of the cytotoxic drug depends on the molecular target and the chemical structure of the conjugate. The drug is expected to exert a potent cytotoxic activity after internalization via receptor-mediated endocytosis, which means that a good membrane permeability is needed to diffuse across the endosomal membranes and some organelles (e.g. mitochondrion or nucleus). In addition, its chemical properties must afford an easy conjugation to the linker or self immolative moiety, for example through functional groups like hydroxyl, carboxyl, amines, carbonyls or thiols. In some cases it is also necessary to regulate the hydrophobicity of the payload to improve the conjugate’s pharmacokinetics, this can be achieved by introducing hydrophilic PEG spacers.^18,50

Among the cytotoxic drugs used in SMDC currently in clinical trials there are antimitotic agents such as paclitaxel, docetaxel, tubulysin B and desacetyl vinblastine hydrazide;

and DNA damaging agents like ifosfamide and mitomycin C (Figure 13).⁵⁰ However, promising cytotoxic agents with higher potency (IC50 < 10^-9M) are also being evaluated as payloads for SMDCs. This is the case of the tubulin polymerization inhibitors monomethyl auristatin E (MMAE),^58,73 cryptophycins⁷⁴ and maytansinoids (e.g.

DM1),^57,75 and the intercalating agent PNU-159682.⁷³

(26)

25 Figure 13. Cytotoxic drugs used in SMDCs

1.4.5. SMDCs in clinical trials

At present there are nine SMDCs in different stages of clinical trials (Table 1).^47,50 Most of these conjugates target the vitamin folate receptor alpha (FRα) overexpressed in a variety of human tumors (e.g. ovary, lung, kidney, endometrium, colon and breast) and have been developed by Endocyte. Their lead candidate EC145, also known as vintafolide, contains the vitamin folic acid conjugated to desacetyl vinblastine monohydrazide (DAVBH) via a dithiol cleavable linker. It is currently in Phase II testing against non-small-cell lung carcinoma and solid tumors.⁷⁶

Table 1. Small molecule-drug conjugates in clinical trials^50,77

Conjugate Target Ligand Cytotoxic agent State

Glufosfamide Glut1 β-D-glucose Ifosfamide

Phase 3: metastatic pancreatic cancer Phase 2: GM, pancreatic cancer and soft tissue sarcoma

Phase 1: pancreatic neoplasm

(27)

26

NGR-TNF APN NGR TNF-α

Phase 2: metastatic ovarian cancer, metastatic SCLC, adult soft tissue carcinoma, metastatic HCC, colorectal cancer Phase 1: advanced solid tumors

GRN1005 LRP1 Angiopep2 Paclitaxel

Phase 2: brain cancer, glioma, metastatic breast cancer

BIND 014 PSMA DUPA Liposomal

docetaxel

Phase 2: prostate cancer, NSCLC

EC1169 PSMA DUPA Tubulysin Phase 1: recurrent

MCRPC

EC145 FR Folic acid

Desacetyl vinblastine

hydrazide (DAVBH)

Phase 2: solid tumors, ovarian and endometrial cancer, NSCLC, lung

adenocarcinoma Phase 1: recurrent or refractory solid tumors

EC0225 FR Folic acid DAVBH and

mitomycin C

Phase 1: refractory or metastatic solid tumors

EC0489 FR Folic acid DAVBH

Phase 1: refractory or metastatic solid tumors

EC1456 FR Folic acid tubulysin

Phase 1: solid tumors, NSCLC, ovarian cancer Glut1 = glucose transporter 1, TNF = tumor necrosis factor, APN = aminopeptidase N, LRP1 = low-density lipoprotein receptor-related protein 1, PSMA = prostate-specific membrane antigen, DUPA = 2-[3-(1,3- dicarboxypropyl)ureido]pentanedioic acid, DAVBH = desacetylvinblastine hydrazide, GM = glioblastoma multiforme, SCLC = small-cell lung cancer, NSCLC = nonsmall-cell lung cancer, HCC = hepatocellular carcinoma, FR = folate receptor, MCRPC = metastatic castration-resistant prostate cancer.

(28)

27 Endocyte has also developed a SMDC targeting the prostate specific membrane antigen (PSMA). EC1169 is constituted by the ligand DUPA and tubulysin B hydrazide as a payload, and it is currently in Phase I clinical trials for recurrent metastatic castration- resistance prostate cancer (MCRPC) patients.^78,79 Another SMDC targeting the PSMA is BIND 014, developed by Bind Therapeutics. This conjugate containing DUPA and a docetaxel-based nanoparticle payload, has recently shown acceptable safety in phase II with MCRPC patients.⁸⁰

NGR-TNFα is a SMDC developed by Corti and coworkers⁸¹ that targets the aminopeptidase-N (APN), overexpressed on the vasculature of several human cancers.

It contains a small peptide ligand formed by the sequence Asn-Gly-Arg, conjugated to the tumor necrosis factor α (TNFα), a multifunctional cytokine that plays a key role in apoptosis and cell survival, as well as in inflammation and immunity.⁸² This compound has shown promising activity in Phase I clinical trials in combination with doxorubicin for treatments of solid tumors.⁸¹ Other peptide-drug conjugate in clinical evaluation is GRN 1005 (or ANG 1005), developed by Angiochem and containing three molecules of paclitaxel bound to the 19-aminocid sequence angiopep2 that targets the low-density lipoprotein receptor-related protein 1 (LRP-1).⁸³ This conjugate has been recently tested in non-small cell lung cancer patients with brain metastasis (GRABM-L).

The progress of SMDCs in the pipeline of new therapeutics reassures the validity of this approach. Furthermore, the increasing diversity of ligands, spacers, linkers and payloads now available makes possible the design of novel and more efficient conjugates with improved affinity for their specific targets and better control of drug delivery and release.^9,50

(29)

28

Chapter 2. Tumor targeting with integrin ligands

2.1. Role of integrins in cancer

Integrins arecell surface receptors constituted by α and β subunits associated in a non- covalent manner. Both subunits are type I transmembrane glycoproteins that contain an extracellular domain, a single transmembrane domain and a short intracellular tail. In vertebrates, 18 α subunits and eight β subunits form 24 integrin αβ heterodimers expressed in different tissues.^84,85 Each integrin exhibits a distinct binding affinity to particular ligands determined mostly by the α subunit (Figure 14), defining integrin subfamilies with specificity for Arg-Gly-Asp (RGD) motifs (αIIb, αV, α5, α8), intercellular adhesion molecules and inflammatory ligands (α4, αL, αM, αX, and αD), collagens (α1, α2, α10, α11) and laminins (α3, α6, α7).⁸⁶

Figure 14. Schematic representation of two integrins subunits and the 24 members of integrin family⁸⁶

Integrins play an important role in cell adhesion (cell-cell, cell-ECM), migration, survival and growth. After binding to ECM proteins (e.g. fibronectin and vitronectin) or cell surface immunoglobulin proteins (e.g. ICAM-1 and VCAM-1), integrins initiate a signaling cascade that can include tyrosine phosphorylation of focal adhesion kinases (FAK) and interaction with growth factors receptors (GFRs).^87,88 They mediate a bidirectional “outside-in” and “inside-out” signaling across the cell membrane,

(30)

29 exchanging information between the ECM and intracellular molecules.^89,90 These complex signaling pathways allow the control of cell polarity, mobility, ECM remodeling and assembly that results in cell survival and proliferation.⁹¹ During the last two decades, integrins have gained increasing attention in pharmacological research since some integrins, notably αVβ3, α5β1 and αVβ6, are overexpressed in a variety of cancers contributing to progression and metastasis (Table 2).^92,93

Table 2. Integrins in cancer progression⁹²

Tumor Type Integrin expressed

Melanoma αVβ3, α5β1

Breast αVβ3, α6β4

Prostate αVβ3

Pancreatic αVβ3

Ovarian αVβ3, α4β1

Cervical αVβ3, αVβ6

Glioblastoma αVβ3, αVβ5

Non-small-cell lung carcinoma α5β1

Colon αVβ6

The integrin receptor αVβ3, first identified by Ruoslahti and coworkers,⁹⁴ is widely expressed on blood vessels of tumor cells (e.g. breast, glioblastoma, ovarian, prostate cancer) but not on vessels of normal tissue.⁹⁵ Integrin αVβ3 is upregulated during tumor angiogenesis due to the stimuli of angiogenic growth factors such as fibroblast growth factor-2 (FGF-2), tumor necrosis factor α (TNF-α) and interleukin 8 (IL-8) present at wounds and inflammation sites. This is a critical step in tumor progression and metastasis because it provides oxygen and nutrients to the cells.^96,97 Moreover, the activation of αVβ3 also facilitates tumor cell migration by regulation of the matrix- degrading protease MMP2 on the surface of angiogenic blood vessels, resulting in collagen degradation and ECM modification.^87,98

Because its implication in biological functions determinant for cancer progression and its high expression in tumor tissues, αVβ3 has been largely studied and validated as a therapeutic target.

(31)

30

2.2. Integrin ligands targeting the α

V

β

3

receptor

2.2.1. RGD integrin ligands 2.2.1.1. RGD recognition motif

In 1987, Ruoslahti reported the tripeptide sequence RGD (Figure 15) as the basic motif present in many natural ligands of the αVβ3 receptor such as fibronectin and other cell adhesion proteins (e.g. vitronectin, osteopontin, collagen).⁹⁹

Figure 15. RGD sequence

The complete understanding of the interactions between RGD and αVβ3, however,was only possible in 2002 when Xiong and coworkers¹⁰⁰ reported the crystal structure of the extracellular segment of the αVβ3 integrin receptor complexed with the αVβ3 integrin binder Cilengitide (Figure 16).

Figure 16. Cilengitide–integrin aVb3 interaction. A) Surface representation of the RGD ligand-binding site; (B) Crystal structure of the aVb3 -Cilengitide complex. (ADMIDAS and MIDAS regions are shown in violet and cyan respectively; Cilengitide is represented in yellow; αV and β3 residues are labelled in blue and red, respectively; O and N atoms are represented in red and blue respectively. Hydrogen bonds and salt bridges are represented with dotted lines). Adapted with permission from Jian-Ping Xiong, Thilo Stehle et al.; Science. 2002, 296 (5565), 151–155. Copyright © 2002 The American Association for the Advancement of Science.

(32)

31 The crystal structure showed an extended conformation of the RGD sequence in the binding pocket, with a distance of 9 Å between the Cβ atoms of the Arg and Asp residues. This folding allows the guanidine group of Arg to interact with two anionic aspartic acid residues in the α-subunit (Asp 218 and Asp150), whereas the aspartic acid binds to Mn²⁺ divalent cation in the metal ion-dependent adhesion site (MIDAS) region of the β-subunit. The glycine residue, at the interface between both subunits, presents weak hydrophobic interactions with the carbonyl group of Arg216, same than the aromatic group of the ligand with Tyr122.^100,101 All these integrin-ligand interactions, also known as “electrostatic clamp”, suggested structural requirements that constituted the starting point for the development of high affinity synthetic αVβ3 integrin ligands.

2.2.1.2. Cyclic RGD integrin ligands

Several small molecules, peptides and peptidomimetics containing the RGD sequence have been designed to target the integrin ανβ3 either as antagonists or as vehicles for selective delivery of drugs and imaging probes to tumors.^102,103

First synthetic RGD ligands were linear peptides that included the RGD-motif and other amino acids added to the sequence (e.g. RGD, RGDS, GRGD, GRGDS, GRGDSP, GRGDSPK). Whereas linear ligands showed good binding affinity for the ανβ3 receptor, also presented low stability regarding enzymatic degradation, resulting in limited applicability for in vivo studies.^104,105 This led to the use of different strategies including cyclization, modification of the stereochemical configuration of the constituent amino acids and N-methylation to improve the biological activity of RGD ligands.^106,107

In 1991, Kessler and coworkers developed the cyclic pentapeptide cyclo(RGDfV) (Figure 17a) which displayed high binding affinity towards ανβ3 (IC50 ανβ3 = 1.54 ± 0.12 nM) while retaining selectivity against other integrins (e.g. αIIbβ3, ανβ5, ανβ8).^108,109 This base structure was later modified to produce new ligands with improved activity and selectivity profiles.¹¹⁰ Among them, cyclo[RGDf-(NMe)V] (Cilengitide, Figure 17b) showed outstanding binding affinity for ανβ3 (IC50 ανβ3 = 0.61 ± 0.06 nM) and ανβ5 (IC50

ανβ5 = 8.4 ± 2.1 nM) as well as subnanomolar antagonistic activity for the ανβ3

receptor.^111,112 Cilengitide became the first integrin antagonist to be tested in clinical trials and it is currently undergoing phase II studies for the treatment of different

(33)

32 tumors^113–115 despite its failure in phase III trial for the treatment of patients with newly diagnosed glioblastoma.¹¹⁶

Figure 17. Structure of cyclic RGD ligands c(RGDfV) and c[RGDf-(NMe)V]

Other RGD ligands reported by the same group are cyclo(RGDfK), cyclo(RGDyK), cyclo(RGDfC) (Figure 18, a-c) and RGD4C (Figure 18, d).^95,103 The amino group of the lysine residue of cyclo(RGDfK) and cyclo(RGDyK) allows further chemical conjugation, similarly in the case of cyclo(RGDfC), the thiol group of the cysteine residue is often used in the conjugation to maleimide-functionalized linkers via Michael addition. For these reason, the mentioned ligands are commonly used as vectors for the delivery of therapeutic agents.^95,117,118 In 1995, Ruoslahti and coworkers reported the discovery by phage display technology of the undecapeptide RGD4C (ACDCRGDCFCG)¹¹⁹, which is structurally constrained by two disulfide bonds. RGD4C has been used delivery systems by conjugation at its N- or C-terminals.^120–123 Furthermore, it can be expressed by recombinant methods into proteins and viruses, as in the case of the RGD4C-TNF fusion protein,¹²⁴ used for the targeted delivery of TNF to ανβ3 expressing tumors.

Figure 18. RGD ligands used in delivery systems⁹⁵

(34)

33 2.2.1.3. Cyclo[DKP-RGD] integrin ligands

In 2009 our research group reported the synthesis of low-nanomolar peptidomimetics where the RGD sequence was cyclized by means of a bifunctional diketopiperazine (DKP) scaffold containing a carboxylic and an amino function.^125,126 In 2012, a small library of different cyclo[DKP-RGD] ligands was reported, differing in the configuration at the two DKP stereocenters (position 3 and 6) and in the substituents at the DKP nitrogen atoms (Figure 19). The introduction of the DKP scaffold confers metabolic stability and conformational rigidity to the ligand, facilitating the interactions needed to fit into the RGD pocket of ανβ3 receptor.^127,128 In particular, the RGD peptidomimetics 2- 7 derived from trans-DKP scaffolds (DKP2-DKP7) showed a preferential binding affinity towards integrin αVβ3, inhibiting the binding of biotinylated vitronectin, a natural integrin ligand, to the purified αVβ3 at low-nanomolar IC50 values in a competition binding assay (Table 3).¹²⁷

Figure 19. Library of cyclo[DKP-RGD] integrin ligands

Table 3. Inhibition of biotinylated vitronectin binding to ανβ3 and ανβ5 receptors^126,127 Compound N° Structure ανβ3 IC50

[nM]

ανβ5 IC50 [nM]

1 cyclo[DKP1-RGD] 3898 ± 418 >10⁴

2 cyclo[DKP2-RGD] 3.2 ± 2.7 114 ± 99

3 cyclo[DKP3-RGD] 4.5 ± 1.1 149 ± 25

4 cyclo[DKP4-RGD] 7.6 ± 4.3 216 ± 5

5 cyclo[DKP5-RGD] 12.2 ± 5.0 131 ± 29

6 cyclo[DKP6-RGD] 2.1 ± 0.6 79 ± 3

7* cyclo[DKP7-RGD]

a) 220.2 ± 82.3

b) 0.2 ± 0.09

a) >10⁴

b) 109 ± 15

8 cyclo[DKP8-RGD] 7.5 ± 0.0 >10³

(35)

34

- cyclo(RGDfV) 3.2 ± 1.3 7.5 ± 4.8

[a]IC50 values are calculated as the concentration of compound required for 50% inhibition of biotinylated vitronectin binding. * Two diastereoisomers detected.

NMR characterization and conformational studies performed on the cyclo[DKP-RGD]

ligands 1-8 revealed that the ligands with highest affinity values (trans configuration) displayed well-defined preferential conformations with an average distance of 8.8Å between Cβ(Arg) and Cβ(Asp) consistent with an extended arrangement of the RGD sequence (Figure 20).¹²⁷

Figure 20. Structures obtained by restrained MC/SD simulations based on NOESY spectra distance information¹²⁷

Docking studies were performed based on the representative conformations obtained from the MC/SD simulations. During these studies the ligand cyclo[DKP3-RGD] 3, produced top-ranked poses displaying all the important interactions RGD- αVβ3 integrin, taking as a reference the crystal structure of the extracellular segment of the αVβ3

integrin receptor complexed with the peptide binder Cilengitide.¹²⁷

(36)

35 Figure 21. Best pose of compound cyclo[DKP3-RGD] into the crystal structure of avb3 integrin overlaid on

Cilengitide (green tube representation).

In further biological evaluation, the c[DKP3-RGD] ligand 3 was tested for its effect on cell viability, proliferation, migration and capillary network formation; mRNA expression of αV, β3 and β5 subunits and Akt phosphorylation in human umbilical vein endothelial cells (HUVEC). Results showed that cyclo[DKP3-RGD] significantly inhibit the cell adhesion and angiogenesis induced by growth factors (VEGF, EGF, IGF-I, FGF2 and IL-8) as well as the phosphorylation of Akt, a protein kinase important in the regulation of vascular homeostasis and angiogenesis.¹²⁹ Recent studies demonstrated that c[DKP3- RGD] was able to inhibit also the FAK/Akt integrin-activated transduction signaling pathway and integrin-mediated cell infiltration process in U373 human glioblastoma cell line, reinforcing its condition of true integrin αvβ3 antagonist.¹³⁰

The former results increased the interest in cyclo[DKP3-RGD] 3 as a potential vehicle for the delivery of cytotoxic agents. For this purpose the DKP3 scaffold was modified by substituting one amine proton with a benzylamine moiety, obtaining the functionalized c[DKP-f3-RGD] ligand (9, Figure 22)¹³¹ that has been used in the preparation of SMDCs targeting the αvβ3 receptor.^70,132,133

Figure 22. Cyclo[DKP-f3-RGD] (c[DKP-RGD]-CH2NH2) integrin ligand (9)